Colorimetric and Resolution requirements of cameras

Colorimetric and Resolution requirements of cameras
Alan Roberts
Tests and settings on a Canon XF300/305
Tests were made on a Canon XF305E, serial number 263050000050, after preliminary tests on a prototype
model. The camera has 3 1/3” CMOS sensors and an integral lens. It records MPEG2-compressed video,
but all the tests were made using the HDSDI output, captured uncompressed and analysed in software.
The 305 model has HDSDI output, can be genlocked, and has timecode connection, while the 300 has none
of these connections. Both models are available in ‘E’ and ‘A’ versions, working at 50Hz and 59.94Hz
respectively. A hardware upgrade can be installed in either to make the camera work at both rates, but tests
were made only on an unmodified ‘E’ (50Hz, PAL) model. The distinction between models is obscure, and
is apparent only on the makers label underneath the camera. Casual users would not be aware of these
differences.
The camera weighs 2.65kg, which is fairly heavy for a small-format camera. The built-in lens has 18:1 zoom
range from 4.1mm to 73.8mm focal length (equivalent to 29.3 to 527mm in a full-frame 35mm camera). The
sensors are CMOS, full-resolution 1920x1080. Recording is onto Compact Flash card (two slots) in MPEG2, long-GoP, with MXF file format. Three bit rate options are available: 50Mb/s CBR (constant bit rate) at
4:2:2 colour sampling (1920x1080 or 1280x720), 35Mb/s 4:2:0 VBR (1920x1080 or 1280x720) and 25Mb/s
4:2:0 CBR (1440x1080 only). Thus it complies with broadcast requirements for bit rate and offers more
economic rates for greater economy. At these rates, a 64GB card can record 160, 225 and 310 minutes
respectively. In 1080 mode, both interlaced and progressive modes are available. Off-speed recording at
fixed speeds from 12 to 60fps is possible.
There is a conventional viewfinder (approximately 1650x935 pixels), plus a separate LCD screen
(approximately 1480x830 pixels) front mounted, which can be swung out to either side of the camera. Both
displays are adequate for focusing.
There are neutral filters for exposure control, and manual control of the lens. Sensitivity is rather good,
although it is specified in an obscure way. On-screen video level monitoring is good, there are options for
both waveform monitoring and vectorscope. There is an image magnifier as a focus aid.
Connectivity is good, with HDSDI and timecode (only in the 305 model), plus HDMI and USB, analogue
component, BNC analogue video, 3.5mm multifunction jack socket and XLRs for audio. Power
consumption is about 9 watts at 7.2 volts.
The camera performed well under test.
1
Colorimetric and Resolution requirements of cameras
Alan Roberts
Tests and settings on a Canon XF300/305
Camera settings which affect picture quality directly, such as gamma, detail and matrix are held in scene
files. These are available via a small button at the back left-hand side of the camera, adjacent to the menu
button. Control is simple, with a single rotary control and a ‘set’ button.
Many of the menu items have little or no effect on image quality. Those that have significant effect are
highlighted. The full set of menu items is given for completeness. In boxes with a range of numeric settings,
e.g. -99~+99, the values indicate the range, and zero means no alteration to factory setting, not zero effect,
and no scales are given in the manuals. For each item, the factory setting is underlined. “Pref” (preferred)
settings are in the last column, where appropriate, for normal video shooting and for film-look shooting.
Where no preferred value is given, either the factory setting is best, or the setting does not have great effect
on image quality. In some instances, it is possible to alter the menus such that they produce more
meaningful numbers. Menus are nested: items in bold-face in the listing are headings leading to a further
nested menu.
Measurement results are given in section 2, after the menus.
This listing of the menus and contents is complete, but this should not be used as an excuse for not reading
the manuals.
2
1
Menus and settings
CUSTOM PICTURE (i.e. scene file)
Item
Select File
Edit file
Rename
Protect
Reset
Gamma
Black
Master Pedestal
Master Black
Red
Green
Blue
Black Gamma
Level
Range
Point
Low Key Satur
Enable
Level
Knee
Enable
Automatic
Slope
Point
Saturation
Sharpness
Level
H Detail Freq
Coring
Level
D-Ofst
D-Curve
D-Depth
HV Detail Bal
Limit
Select
Knee Aperture
Gain
Slope
Range
Off, C1~C9, SD1~20
description
9 scene files, 20 on SD card.1
text
Unprotect, Protect
Cancel, OK
Normal1, Normal2, Normal3, Normal4,
Cine1, Cine2
2
Main camera settings
Pref
Video
Film
Normal 3
Cine1
-50~0~+50
-50~0~+50
-50~0~+50
-50~0~+50
Use sparingly, affects noise
+ expands blacks, compresses
-50~0~+50
0
-5~0~+50
-1~0~+50
0
0
On, Off
-50~0~+50
Off3
Highlight compression, not
available in Cine gammas
On, Off
On, Off
-35~0~+50
50~95~109
-10~0~+10
On
Off
4
Lovely, IRE values!
Preserve colour in highlights
85
0
0
-10~0~+50
-8~0~+8
-3
+8
To avoid sharpening noise
-30~0~+50
0~50
0~8
-4~0+4
-8~+8
-50~0~+50
0~15
+2
+5
0
Aperture correction
+2
+15
5
0~9
0,1~3
1
Prevent sharpening near
black
Level Depend
________________________________________________________________________________________________
1
By default, files 1~6 are available for editing, 7~9 protected. 7=factory settings for video shown on consumer displays,
7=film-look on decent monitor, 9=video for printing to film. Protection can be removed for over-writing.
2
Gamma curves: Normal 1=NHK 4.0, Normal 2=ITU709 4.0 (i.e. the 709 curve with lower slope near black), Normal
3=ITU709, Normal 4=BBC 0.4. Cine 1=film for video, Cine 2=for transfer to film.
3
Low-key Saturation helps keep colouring right near black, but can worsen noise, use with care.
4
Knee was not explored during the tests, because Cine 1 gamma curve is good for a film look. Use Knee Slope to fine
tune the video gamma curves if needed, point should be left near 85%.
5
Use Knee Aperture gain in conjunction with Knee Slope; as Knee Slope goes down, Aperture Gain can go up, to
maintain sharpness.
3
Level
Slope
Offset
Noise Reduction
Skin Detail
Effect Level
Hue
Chroma
Area
Y Level
Selective NR
Effective Level
Hue
Chroma
Area
Y Level
Color Matrix
Select
Gain
Phase
R-G
R-B
G-R
G-B
B-R
B-G
White Bal
R Gain
G Gain
B Bain
Color Correction
Select
Area A Setting
Phase
Chroma
Area
Y Level
Area A Revision
Level
Phase
Area B Setting
Phase
Chroma
Area
Y Level
Area B Revision
Level
Phase
Others
Setup Level
Level
Press
6
0~50
0~3
0~50
1~8, Off, Automatic
Automatic7
Soften skin tones
High, Middle, Low, Off
-16~0~+16
0~16~31
0~16~31
0~16~31
Fine tune noise reduction8
High, Middle, Low, Off
0~16~31
0~16~31
0~16~31
0~16~31
Normal1, Normal2, Normal3, Normal4,
Cine1, Cine2
-50~0~+50
-50~0~+50
-50~0~+50
-50~0~+50
-50~0~+50
-50~0~+50
-50~0~+50
-50~0~+50
Preset matrices9
Normal 3
Cine 1
-810
-50~0~+50
-50~0~+50
-50~0~+50
Tweak two colours
Off
Off, Area A, Area B, Area A&B
0~31
0~31
0~31
0~31
-50~+50
-50~+50
0~31
0~31
0~31
0~31
-50~+50
-50~+50
-50~0~+50
On, Off
Scales Black and Master Ped
Squeezes video to 100%
Off11
________________________________________________________________________________________________
6
Use Level Depend in conjunction wit Black Stretch, to prevent over-sharpening near black.
7
Very effective noise reduction, see measurements section. Setting to 8 reduces resolution to 1280x720 very cleanly.
8
Selective Noise Reduction could be useful but will take significant time to set up to be effective.
9
Same rules as for gamma curves: Normal 1=NHK 4.0, Normal 2=ITU709 4.0 (i.e. the 709 curve with lower slope near
black), Normal 3=ITU709, Normal 4=BBC 0.4. Cine 1=film for video, Cine 2=for transfer to film.
10
Matrix, see measurements section below.
11
Press, not tested. This could be handy for shoot-and-run operations where exposure control is difficult.
4
Clip 100% IRE
Transfer File
Copy To
Load From
Copy To Cam
Load From Cam
Add CP File
On, Off
Clips hard at 100%
SD1~SD20
SD1~SD20
C1~C9
C1~C9
C1~C9
Copy current file to SD card
Copy from SD card to current
Adds file to recording clip
5
Off
CAMERA SETUP
Item
Light Metering
AE Shift
AGC Limit
Gain
L
M
H
Fine Tuning
White Balance
Range
Backlight, Standard, Spotlight
+2, +1.5, +1.25, +1, +0.75, +0.5, +0.25, 0,
-0.25, -0.5, -0.75, -1, -1.25, -1.5, -2
Off (21), 18, 15, 12, 9, 6, 3dB
Fine tuning, 33, 21, 18, 12, 6, 3, 0, -3, 6dB
Fine tuning, 33, 21, 18, 12, 6, 3, 0, -3, 6dB
Fine tuning, 33, 21, 18, 12, 6, 3, 0, -3, 6dB
0~21dB
Daylight, Tungsten, Kelvin
AF Mode
Speed
Instant, Normal
Face AF
On, Off
Focus Limit
On, Off
Image Stabiliser
Zoom
Speed Level
Grip Rocker
Constant Speed
Handle Rocker H
Handle Rocker L
Wireless Controller
Teleconverter
Flicker Reduction
Wide Attach Lens
Color Bars
Enable
Type
Powered, Dynamic, Standard, Off
Fast, Normal, Slow
Constant, Variable
1~8~6
1~16
1~8~16
1~8~16
On, Off
Automatic, Off
WA-H82, Off
Main video standard setting
Pref
description
Auto exposure compensation
Deviation from auto-exposure,
stops
Maximum gain AGC will go to
-6dB
-3dB
6dB
0.5dB steps, wow!
Kelvin lets you set the colour temp.
directly
Auto focus
Manually tweaking the focus ring
overrides auto mode
Auto focus on a face
On sets closest to 1m; off sets to
2cm for macro use
Degree of stabilisation, powered for
tripod use, dynamic for walking,
standard is lowest.
Variable speed is pressure sensitive
Sets constant speeds, 1 (slow)=5
minutes, 1 (fast)=1 minute, 16
(slow)=4.5 sec, 16 (fast)=1.8 sec.
Compensate for lighting flicker
Compensate for wide-angle lens
On, Off
Type 1, Type 2
SMPTE or ARIB
Range
description
Type 112
AUDIO SETUP
Item
Audio Input
XLR Rec CH
Int Mic Low Cut
Int Mic Sensitivity
Int Mic Att
XLR1 Mic Trimming
XLR2 Mic Trimming
XLR1 Mic Att
XLR2 Mic Att
XLR ALC Link
Limiter
1kHz Tone
Audio Output
Monitor Delay
Channel
Level
CH1, CH1/2
Off, LC1, LC2
Normal, High
On, Off
+12, +6, 0, -6, -12dB
+12, +6, 0, -6, -12dB
On, Off
On, Off
Linked, Separate
On, Off
-12, -18, -20dB, Off
Pref
1 for voices, 2 for wind cut
High=+6dB
On=-12dB
Line out, Normal
CH1/2, CH1/1, CH2/2, All/All
1V rms, 2Vrms
Use link for stereo recording
Limit at -4dB13
Tone over colour bars
Does not affect recording
Set headphone sound delay
All does mono mix
1V=0dB, 2V=+6dB
________________________________________________________________________________________________
12
Either bars are acceptable. ARIB bars contain more useful aspects than SMPTE, but SMPTE are widely accepted as
the standard for HDTV.
13
Manual audio controls (knobs): 0=off, 5=0dB, 10=+18dB
6
VIDEO SETUP
Item
SDI Output
Component Output
HD Onscreen Disp
SD Onscreen Disp
SD Output
Range
HD, SD, Off
HD, SD
On, Off
On, Off
Squeeze, Letterbox, Side crop
description
Off saves battery power
Pref
Puts camera screen messages
on output, not on recording
LCD/VF SETUP
Item
LCD Setup
Brightness
Contrast
Color
Sharpness
Backlight
VF Setup
Brightness
Contrast
Color
Sharpness
Backlight
LCD/VF B&W
LCD/VF Simul
Metadata Display
Date/Time
Camera Data
Peaking
Select
Peaking 1
Color
Gain
Frequency
Peaking 2
Color
Gain
Frequency
Zebra
Select
Zebra 1 Level
Zebra 2 Level
HD Output
Markers
Enable
Center
Horizontal
Grid
Safety Area
Select Area
Aspect Marker
Aspect Ratio
Audio Level
Custom Display
Remaining Battery
Range
description
Pref
-99~0~+99
-99~0~+99
-20~0~+20
1, 2, 3, 4
Normal, Bright
-99~0~+99
-99~0~+99
-20~0~+20
1, 2, 3, 4
Normal, Bright
On, Off
On, Off
On=mono
On=v/f and LCD on together
On, Off
On, Off
On, Off
Peaking 1, Peaking 2
Only in playback mode
Two settable peaking regimes
White, Red, Yellow, Blue
Off, 1~8~15
1, 2, 3, 4
White, Red, Yellow, Blue
Off, 1~15
1, 2, 3, 4
On, Off
Zebra 1, Zebra 2, Zebra 1&2
70, 75, 80, 85, 90, 95%
70~100%
Active over ±5% of target
Active above target level14
Show zebras on HD video
outputs
On, Off
On, Off
White, Gray, Off
White, Gray, Off
White, Gray, Off
White, Gray, Off
80, 90, 92.5, 95%
White, Gray, Off
4:3, 13:9, 14:9, 1.66:1, 1.75:1, 1.85:1, 2.35:1
On, Off
Warning, Normal, Off
70
100
14:915
Warning=only when low batt
________________________________________________________________________________________________
14
When zebra patterns overlap. Zebra 1 takes priority. This is perhaps the best use of zebras I’ve found in any camera
yet.
15
Not essential, but useful when shooting for mixed 16:9/4:3 delivery.
7
Remaining Tec Time
Rec Mode
Time Code
Zoom Position
Light Metering
Custom Picture
ND Filter
Focus Ring Warn
Focus Mode
Object Distance
Full Auto
White Balance
Exposure
Iris
Gain
Shutter
Wide Attach Lens
Teleconverter
Peaking
Magnification
Image Stabiliser
Interval Counter
SD Card Status
Bit Rate/Resolution
Frame Rate
Character Rec
Wireless Controller
Output Display
SDI Rec Command
User Memo
Audio Output Ch
Audio Level
Date/Time
Warning, Normal, Off
On, Off
On, Off
Always on, Normal, Off
On, Off
On, Off
Warning, Normal, Off
On, Off
On, Off
Warning, Normal, Off
On, Off
On, Off
On, Off
On, Off
On, Off
On, Off
On, Off
On, Off
On, Off
On, Off
On, Off
On, Off
Warning, Normal, Off
On, Off
On, Off
On, Off
On, Off
On, Off
On, Off
On, Off
On, Off
On, Off
Date, Time, Date/Time, Off
and so on
Normal=only when changing
Normal=only when changing
Normal=only when focusing
Time-code and User Bits
Pref
TC/UB SETUP
Item
Timecode
Mode
Run
DF/NDF
Setting
TC In/Out
User bits
Rec Mode
Output Mode
Type
Range
description
Preset, Regen
Rec run, Free run
DF, NDF
Set, Reset
In, Out
Relevant only in ‘A’ and unlocked ‘E’ cameras16
Opens menu to set TC and UB
Not available on XF300 models
Internal, External
Fixed, Pulldown
Setting, Time, Date
Not available on XF300 models
OTHER FUNCTIONS
Item
Reset All Settings
Transfer Menu
Save To
Load From
Time Zone
Range
Cancel, OK
description
Menu, Menu+CP
Menu, Menu+CP
Load save menus, or menus
and all custom presets
Default for ‘A’ model is New
York (UTC-5.00)
Date and Time
-12.00~+14.00
Clock Set
Date Format
Pref
YMD, YMD/24H, MDY, MDY/24H, DMY,
DMY/24H
________________________________________________________________________________________________
16
DropFrame timecode for recording at ‘NTSC’ speeds (59.94i, 29,97p, 23.98p), is shown as hh:mm:ss.ff,
NonDropFrame shows as hh:mm:ss:ff.
8
WFM (LCD)
Setting
Waveform Monitor
WFM, VS, Edge mon, Off
Line, Line+spot, Field, RGB, YPbPr
Gain
Vectorscope
Gain
1x, 2x
Spot, Normal
1x, 5x
Edge Monitor
Type 1, Type 2
Language
Wireless Controller
Assign Button
Tally Lamp
Front
Rear
Media Access LED
Genlock
Bit Rate/Resolution
NTSC/PAL
Frame Rate
Special Rec
Interval Rec
Interval
Rec Frames
German, English, Spanish, French, Italian,
Polish, Russian, Simplified Chinese, Japanese
On, Off
Camera mode :None, Standard IS, Dynamic
IS, Powered IS, Focus limit, Face AF, Face
select, Backlight, Spotlight, Teleconverter,
Peaking, Zebra, WFM (LCD), Magnification,
Color bars, Marker, LCD setup, LCD/VF
B&W, Onscreen display, Shot mark 1, Shot
mark 2, Add OK mark, Add check mark, Time
code, Time code hold, Audio output CH, Audio
level, Wireless controller, Photo, Rec review,
Delete last clip
Media mode: None, WFM (LCD), LCD setup,
LCD/VF B&W, Onscreen display. Shot mark 1,
Shot mark 2, Add OK mark, Add L mark, Time
code hold, Audio output CH, Audio level,
Wireless controller, Photo
On, Off
On, Off
On, Off
-1023~0000~+1023
50Mb/s 1920x1080, 50Mb/s 1280x720, 35Mb/s
1920x1080, 35Mb/s 1280x720, 25Mb/s
1440x1080
NTSC, PAL
60i, 60p, 30p, 24p, 50i, 50p, 25p
50i, 50p, 25p
Interval rec, Frame rec, Pre rec, Slow & fast
motion, Off
1~10, 15, 20, 30, 40, 50 sec, 1~10 min
1, 3, 6, 9
2, 6, 12
2, 6, 12
Edge monitor not available in
Playback mode
Spot adds waveform for the
screen area in the red frame
+6dB gain
+14dB gain
Focus aid, shows frequency
content
Language for screen
messages. Menus/settings
remain in English
Remote control
13 (!) assignable buttons,
marked:
1=IS
2=Peakin
3=Zebra
4=WFM
5=Return
6=Magn (right)
7=Magn (top)
8-13 playback buttons
Horizontal phase, XF305 only
50Mb/s
1920x108017
Only in unlocked models
‘A’ NTSC
Both models
when unlocked
‘E’ PAL
Various non-standard
shooting modes
60i or 30p
60p or 24p
‘A’ NTSC
‘E’ PAL
Frame Rec
Rec Frames
1, 3, 6, 9
2, 6, 12
2, 6, 12
60i or 30p
60p or 24p
‘A’ NTSC
‘E’ PAL
Slow & Fast Motion
Rec Frame Rate
Clips
Title Prefix
Number Setting
Delete Last Clip
Copy All Clips
Copy OK Clips
Delete All Clips
Delete All OK Marks
12, 15, 18, 20, 21, 22, 24, 25, 26, 27, 28, 30, 32,
34, 36, 40, 44, 48, 54, 60
12, 15, 18, 20, 21, 22, 24, 25, 26, 27, 28, 30
AA~ZZ
Set, Reset
Cancel, OK
Cancel, OK
Cancel, OK
Cancel, OK
Cancel, OK
50 or 35Mb/s
24Mb/s
Variable
speed
shooting
Text entry
Copy clips card to card
Copy only OK-marked clips
Delete all except OK-marked
Un-mark all cips
________________________________________________________________________________________________
17
This is the minimum coding specification accepted for EBU broadcasting at the time of writing (June 2010).
9
Rec Review
Set Metadata
User Memo
Country Code
Organization
User Code
SDI Rec Command
Photo Numbering
Add CP File
To Clip
To Photo
Delete All Photos
Custom Function
Shockless Gain
Shockless WB
AE Response
Iris Limit
I. Ring Direction
F.Ring Control
F. Assist B&W
Obj. Dist Unit
Zoom Indicator
ZR-2000 AE Shift
Scan Reverse Rec
Character Rec
Reset Hour Meter
Initialization
CF A
CF B
SD Card
Firmware
Entire clip, last 4 sec
Play last 4 seconds of last clip
Off, select from files
4 letters
4 letters
4 letters
On, Off
Reset, Continu
Requires extra software
Entre label, A~Z, 0-9 +-; and
space
Only on XF305
On, Off
On, Off
OK, Cancel
Copies settings to clip or
photo as metadata
Wipes the SD card
Fast, Normal, Slow, Off
On, Off
Fast, Normal, Slow
On, Off
Reverse, Normal
Fast, Normal, Slow
Both, Magnify, Peaking, Off
Meters, Feet
Bar, Number
AE shift, Iris
On, Off
On, Off
Cancel, OK
Auto gain control
Auto white tracking
Card formatting
Cancel, OK
Cancel, OK
Complete, Quick
Shows firmware versions
10
2
Measurement results
2.1 Colour performance
Assessments were made visually, using Macbeth charts as usual. Performance was generally good, but the
skin-tone colours were rather pink. Using the matrix, it was possible to effect some improvement by setting
the R-G value to -8, and there may have been more improvements to be gained by spending more time on
this assessment, probably setting R-G to a negative value as well, and possibly using the colour corrector.
However, the results from the brief test session, using just R-G were quite pleasing and acceptable.
2.2 Gamma curves
There are 4 normal gamma curves available in the camera, and two cine curves. Gamma 1 was clearly the
intended factory setting, and initial tests were made using it, but later examinations established that Gamma
3 is the ITU-709 curve and Gamma 4 the BBC 0.4 curve. For broadcast purposes either of these curves is
acceptable. Although the BBC curve always produces more accurate colour rendition, the 709 curve is
normal for HDTV shooting, so all further tests used Gamma 3.
Experiments with the Knee function established that the camera has about 100% (1 stop) of exposure
headroom. While it was perfectly possible to derive settings which would exploit this using standard gamma
curves and the knee, it is probably better to use one of the cine gamma curves to achieve a film look. This
will be investigated further when a production model is available for test.
2.3 Resolution
A HDTV zone plate chart was used. This contains six circular patterns that fully explore the spatial
frequency performance of the camera, up to 1920x1080 pixels per width and height. There are patterns for
grey-scale testing of luma performance, the others are coloured for examining chroma resolution or other
colour filtering. Modulation is cosine rather than square wave. Each pattern is a “phase space” map of the
possible frequencies that the camera can be expected to deal with, reaching 1920 pixels/picture width (960
cycles) horizontally, and 1080 lines/picture height (540 cycles) vertically.
2.3.1 Resolution, 1080psf
Fig.1 shows a single quadrant of one pattern; for this
exposure, the camera detail enhancement was turned
down to minimum level (-10) which presumably
means no correction, so this is probably the native
performance of the camera. There are clearly no null
zones, where the wanted lower frequencies mix with
aliases produced by spectral folding of the unwanted
higher frequencies, alias products. This is good
evidence that the camera has 3 sensors of full
1920x1080 resolution, and that a “quarter-wave” filter
(bi-refringent crystal or other fabricated filter) has been
included in the optical path, and is well suited to the
camera’s resolution.
There was no evidence of aliasing caused by out-ofband frequencies in the coloured zone-plate patterns.
Usable resolution up to about 900 lines vertically and
1800 horizontally is clear. Also, there was no evidence
through aliasing that “precision offset” (the spatial
offsetting of the green sensor from red and blue by
exactly ½ pixel spacing to improve luma resolution) is Figure 1, resolution, 1080psf, minimum detail (-10)
used in this camera.
11
It is rare to see resolution so clearly as this, which is very encouraging.
2.3.2 Detail enhancement, 1080psf
Figure 2 shows the same quadrant with the detail level
set to maximum (+50). There are new null zones
visible at 2/3 horizontal and vertical resolution limit,
which are due to the small deviation of the camera
gamma curve from a pure power law (this is third
harmonic distortion, emphasised by the detail
enhancement), which is a perfectly normal
phenomenon.
No new aliases have been produced, but there is a
granularity in the higher frequencies caused by the
emphasis of video noise. Clearly, this is not a sensible
setting value to use, and the camera’s ‘zero’ setting is a
good compromise, as is shown in Figure 3. The null
zones at 2/3 limiting frequencies are just starting to be
visible, but the level is quite acceptable.
Figure 2, resolution 1080psf, maximum detail (+50)
Figure 3, resolution 1080psf, factory detail (0)
Figure 4, resolution 1080i, factory detail (0)
2.3.3 Resolution, 1080i
Figure 4 shows the result of setting the camera to interlaced scanning, with the factory detail level setting.
Some vertical detail has been lost as is to be expected from the line-averaging process normally used for
deriving interlaced fields, but it has been lost in a clean way, there are no surprises here.
2.4 Video Noise Levels and Sensitivity
Video noise was measured by recording a white card, uniformly lit, and performing numerical analysis in
software. The camera was set to +12dB gain to ensure a high noise level for measurement, and the results
have been compensated to remove the effect of this higher gain. A high-pass filter was used to remove all
horizontal frequencies below about 5% of the nominal maximum of half-sampling frequency.
12
Two sets of results were obtained, for the camera in default setting, with video noise reduction set to
Automatic, and with noise reduction fully off. The noise levels were a little disappointing for a broadcast
camera (typically -45.5dB), but are typical for a camera with full-resolution HDTV sensors of 1/3 inch size,
and there is little that can be done about it. This value confirms the impression from visual checks made
during the test session
-37
0.0
0.2
0.4
0.6
0.8
-37 0.0
1.0
0.2
0.4
0.6
0.8
1.0
-39
R
-39
-41
-43
R
G
G
B
-41
B
Y
Y
-45
-43
-47
-45
-49
-47
-51
Figure 5, (a) video noise, NR Auto
(b) video noise, NR off
There is an advantage of about 1.5dB to be gained from using the noise reduction in Automatic mode.
However, it should be borne in mind that this advantage is when the camera is set to +12dB gain, and that the
advantage will probably be rather less at 0dB gain. Nevertheless, the noise reduction is worth having.
To investigate the noise reduction process itself, exposures were made of the zone plate test chart, at +12dB
gain. Figure 6 shows the resolution loss in Auto mode.
Figure 6, (a) NR Auto, +12dB
(b) NR off, +12dB
Clearly, little is lost in this mild setting of noise reduction, although some horizontal softening is apparent
but not critical. Figure 7 shows the effect of setting the noise reduction to manual, at level 8 (maximum).
Here, the resolution loss is dramatic, but still there is no introduction of spurious aliasing. The resulting
resolution appears to be nicely limited to a little lower than 1280x720, indicating that the noise reduction
process uses both horizontal and vertical filtering, and is done rather well.
13
Somewhat surprisingly, this control may well be useful
as a pre-filter for shooting footage intended for delivery
at 1280x720, or for down-conversion to 720x576
SDTV. It should produce results considerably superior
to those obtainable through most software downconversion algorithms, or in-camera down-conversions,
because the troublesome higher frequencies are
removed at source.
This is an unexpected benefit, which could be of great
effectiveness.
Sensitivity was not measured directly.
In broadcast
cameras, the normal specification would be the aperture
setting required to produce full amplitude video when a
90% reflectance card is illuminated with 2000 lux and
the camera is set to 0dB gain and normal shutter (e.g.
1/50). The specification claims minimum illumination
to be 0.08 lux at 60i, with +33dB gain and ¼ second
shutter, and the lens, presumably, wide open (F/1.6).
Compensating for the long shutter gives a light-level Figure 7, NR maximum (manual, 8)
factor of 50/4=12.5, resulting in 1 lux at 1/50.
Compensating for the video gain gives another factor of 1033/20=44.67, resulting in 44.67 lux at 0dB and 1/50.
Deriving a further compensation for the open lens to produce an aperture value for 2000 lux illumination
finally gives a value of F/1.6*√(2000/44.67)=F/10.7. So, the sensitivity, expressed in conventional terms is
F/10.7 for full video at 2000 lux with standard gain and shutter. This is a truly remarkable figure for such a
small sensor.
Perhaps the designers would have been better advised to reduce the head amplifier gain by about 6dB,
resulting in a sensitivity of F/9.6 and returning noise levels 6dB lower (around -50dB).
2.5 Shuttering
The camera has 3 CMOS sensors, and thus can be expected to exhibit the effects of a rolling shutter. A
motion sequence was recorded, of a white card being moved vigorously back and forth in front of a focus
chart. Figure 8 shows two fields from that sequence, which clearly show the expected ‘leaning verticals’
brought about through the rolling shutter (since the top of the field or frame is exposed significantly before
the bottom of the field or frame). For this test, the camera was set to interlace mode, underexposed, and the
shutter effectively turned off (i.e. 1/50 second). Had the shutter been set to a shorter period, the edge
blurring would have been much less, but the slope would have remained.
Figure 8, rolling shutter effect, (a) left to right motion
(b) right to left motion
More spectacularly, the effect of the rolling shutter can be demonstrated by a rotating fan. Figure 9 shows
images recorded of a small fan, two bladed, with symmetrical blades. The camera was et to a short shutter
(about 1/1000, not critical) and the fan speed was adjusted to one of several critical speeds at which a
stroboscopic effect was observed. The blades are compressed when on the left (going up, against the rolling
shutter), grossly expanded on the right (going down, overtaking the rolling shutter).
14
Figure 9, rolling shutter effect, rotating fan, 1/1000 shutter
At one point, the right-hand blade in this sequence of
adjacent recorded frames actually falls apart.
If the shutter is set to a more sensible speed (Figure 10),
then the effect is far less visible, but the blurred blades
still appear to be asymmetrical, being swept to the left.
The combination of sloping verticals and distorted
rotating elements may not often occur in programmemaking, but the effects can be disturbing when they
accidentally occur. This camera is neither better nor
worse than other cameras with CMOS sensors, the
effects are an inevitable consequence of the use of rolling
shutter in then camera.
2.6 Conclusion
This camera performs well at HD, for such a smallimage format. Resolution is very well maintained and is Figure 10, rotating fan, 1/100 shutter
refreshingly alias-free. Detail controls work well, and
the factory settings are good. Noise levels are typical for 1”/3 sensors, but sensitivity is unusually good.
Operating the camera at significantly lower gain will reduce the noise level without sacrificing significant
sensitivity.
The integral lens has a maximum aperture of F/1.6, unusually large for a small camera, and there was mo
perceptible loss of resolution through iris diffraction until the lens was stopped down to F/8, at which point
significant resolution was being lost. Again, this is unusual for such a small image size. Thus, the camera
has a useful aperture range from F/1.6 to about F/6.8. This, together with the 3-stage neutral density filters
(each providing a further 2-stops of control), means that the camera has a much better exposure control range
than is normal in a small camera.
15